The problem is the delivery mechanism. Charging is always going to be a bottleneck, unless we invent mass production supercapacitors on the order of what graphe

The problem is the delivery mechanism. Charging is always going to be a bottleneck, unless we invent mass production supercapacitors on the order of what graphene promised (which is seems to be highly unlikely). Edit: even with supercapacitors, the amperage required seems unlikely. Assuming self-parking as a baseline (although it's not strictly required), cars would be able to position themselves over an automated battery pack replacement station. Replacing the discharged pack could take less time than a petroleum refuel, so EVs could be more convenient than petroleum vehicles (any replacement technology should offer more benefits than what it replaces). The station could then move the pack underground for recharging. This underground area would have much higher density compared to cars (where batteries are, what, 10% of the volume?). In addition, dead packs could be identified and scheduled for pickup and recycling. The first problem that most people would bring up is: replacing a worn pack with a new one. This problem actually turns out to be an advantage: disclose the mileage of the pack and allow the customer to pay less for it. I don't need a 300mi pack if I'm just going 5mi to work. This way, more useful lifetime can still be extracted from old packs while turning the recycling problem into a logistically simpler problem (no more complicated than delivering petroleum to fuel stations). The only real problem with this is the clamping mechanism between the car and the pack. It would need to be standardized. Reliability: mechanical systems fail - it's the undercarriage of the car, there's dust and other particulate getting in the way.